18 | cybersecurity researcher | AI AGENTS | buildinpublic | weekend projects

Joined May 2026
As part of this weekend project I explored Ghidra and Researched about Jev for SOC analyst.
1
11
Back to weekend projects.. Made LivePulse LivePulse takes raw live ICMP traffic, measures how the network is behaving, converts those measurements into behavioral signals, classifies the observed behavior, and displays the reasoning in real time. github.com/SAKSHI-JADIYA/Liv…
4
Learning NETWORKING
4
Yesterday.. Made Weekend project 4:- Air gap call. That alerts admin when a prompt is injected in agent or rest method is invoked by calling ands lock down data in single keypress !! Git hub repo:- github.com/SAKSHI-JADIYA/Air…
1
11
Skipped building this weekend. Went deep on LangChain (LCEL, parsers), RAG, Agentic workflows (ReAct, CrewAI) & AI Ethics. Brainstorming 1 problem statement though.
1
16
AX is the new UX. Software was built for humans clicking, now it gets rebuilt for agents. Stripe just paid $8B for OpenRouter betting the main user of the internet is an agent. Rebuilding all that software is 1000+ new companies. The most obvious opportunity in a long time!
191
110
43
1,368
125,014
#BuildinPublic Learning AI AGENTS. But from the Lens of secure builder Analysing Threats and defenses alongside. Cybersecurity researcher you know!😉
1
15
Since I started reading blogs & articles, I realized they are a goldmine for deeply understanding different perspectives. Today's read:- Building AI AGENT without framework. pub.towardsai.net/building-a…
1
19
Anthropic dropped 3 new Skills - import-memory, morning, and skill-creator. ⭐ By keeping skills modular, they're making the system more stable and letting developers combine them however they want.
1
18
Explored Docker 🐋 and LangChain 🦜 🔗 Getting one step ahead in AI AGENT developer journey. Also Exams coming up Sunday done and dusted.
1
17
This weekend I explored mem0 python library & it's GitHub repo, qdrant vector database. Giving memory to FRIDAY 🎉. So now Friday can remember you :)
1
21
I researched how Venice AI claims to be unrestricted & secure. Its architecture pairs open weights with a 4-tier privacy stack. Using client-side E2EE & local state storage, prompts are decrypted *only* inside hardware-isolated TEEs, completely blinding the telemetry loop. Cool!
1
37
Weekend Build Series: Ep 3! Built "Friday" — a local-first, AES-256 encrypted voice assistant. Flow: Mic → Wake-word filters → Llama 3.1 via Ollama → Offline reply. Zero cloud leaks. Stack: Python, Ollama, Cryptography, Pyttsx3 Code 👇 github.com/SAKSHI-JADIYA/FRI…
1
29
Weekend 2 🚀: Built a modular Web Scraper + Parser from scratch. Crawls sites → cleans DOM → pipes context to local Llama 3.1 → gets structured LLM response. Stack: Python, Streamlit, Selenium, BS4, LangChain, Ollama Code 👇 github.com/SAKSHI-JADIYA/ai-… #BuildInPublic
1
48
Building projects every weekend.🪄 Weekend 1: Netflix 📊 CSV → Cleaned → MySQL → Tableau Dashboard Stack: Python, Pandas, SQL, Batch Processing, Tableau, Git Code: github.com/SAKSHI-JADIYA/net… #BuildInPublic #DataEngineering #SQL
1
2
37